WesTrack was the Federal Highway Administration's (FHWA) test facility in Nevada for

developing performance-related specifications for hot-mix asphalt pavement construction. It also provided some of the earliest data on the performance of Superpave asphalt
mixture designs under high rates of heavy truck loading. When Superpave-designed test
sections placed at the track in June 1997 had very rapid rutting failures, the highway community was concerned that the mixture design and construction procedures might be
missing important, but unknown, constraints. A forensic team composed of academicians,
asphalt industry representatives, and State highway agency engineers was assembled to
study the early failures and, if appropriate, to make recommendations for revising the
Superpave procedures. Their examination of the failures resulted in Report No. FHWA-RD99-134, Performance of Coarse-Graded Mixes at WesTrackPremature Rutting, which is available from FHWA or on the Internet at the Turner-Fairbank Highway Research Center
homepage at www.tfhrc.gov.
During the team's investigation, its members concluded that the asphalt paving community needed a good guide on the design of Superpave mixtures. Such a guide would supplement existing specifications and supporting literature and would incorporate the experience of engineers across the country, including the WesTrack designers, in the initial
years of Superpave mixture design and placement. It would be a useful companion to the
National Asphalt Pavement Association's Superpave Construction Guidelines. This publication, Superpave Mixture Design Guide, was prepared by the forensic team. Its contents are
the views of the team and do not necessarily reflect the views of the U.S. Department of
Transportation.
Note that this version of the guide is not expected to be the final word on Superpave mixture design. Both current research studies and additional field experience are likely to
yield refinements in the future.

T. Paul Teng, P.E.

Director, Office of Infrastructure
Research and Development

Notice
This document is disseminated under the sponsorship of the Department of Transportation in
the interest of information exchange. The United States Government assumes no liability for its
contents or use thereof. This report does not constitute a standard, specification, or regulation.
The United States Government does not endorse products or manufacturers. Trade and manufacturer's names appear in this report only because they are considered essential to the object of
the document.

SUPERPAVE MIXTURE
WesTrack Forensic Team Consensus Report
February 2001
Washington, DC

Introduction
Superpave design methods and tools are
being implemented by many State agencies
to replace the Marshall and Hveem design
methods. In 1999, 2,515 projects, specifying
some 73 million metric tons of Superpave,
were let.[1] The majority of the projects in
1999 and in previous years were constructed
with little or no difficulty. On several projects, there were some problems during this
initial implementation. For the most part, the
causes of the problems have been identified
and have been solved. In 2000, estimates
were that more than 3,900 projects, specifying some 134 million metric tons of
Superpave, would be let; this would represent 62 percent of the total hot-mix asphalt
(HMA) tonnage expected to be contracted for
by State agencies during 2000 in the United
States.[1] Superpave has become the mixture
design method of choice by most State transportation departments across the country.
This document, intended as a companion to
the National Asphalt Pavement Association's
(NAPA) Superpave Construction Guidelines,[2]
is a guide for the HMA designer to maximize
the benefits of Superpave while avoiding
potential problems. The Superpave design
process is part of a total pavement design
system. Superpave is a system of components that work together to provide a
SUperior PERforming asphalt PAVEment. As
traffic levels and loading conditions increase
above 1 million 80-kN (18,000-lb) equivalent
single-axle loads (ESALs) during a pavement's design life, some design areas are not
adequately addressed by the current
Superpave specifications.
This guide discusses several issues that
should be considered during the mixture
design process to maximize the benefits of
this method. The Superpave design process
is documented in publications from the
Federal Highway Administration (FHWA),
the Strategic Highway Research Program
(SHRP), the Asphalt Institute (AI), and the
American Association of State Highway and

Transportation Officials (AASHTO).[3-8] Those

publications should be used for detailed
design information. This guide is intended to
serve as a bridge between existing knowledge and additional tools being developed to
measure and predict Superpave mixture performance under traffic.

This document is a guide for the HMA designer

to maximize the benefits of Superpave while
avoiding potential problems.
A Superpave mix design includes several
processes and decision points. First, design
compaction levels are established and materials are selected and characterized. Then,
mixture samples are prepared and laboratory test results are compared to design criteria. However, the existing Superpave design
system does not properly address performance prediction testing on mixture samples
or decision-making during the design
process. This guide will address both of
these areas.

Superpave Mixture Design

Compaction Level Determination
Prior to 1999, the design ESALs shown in the
Gyratory Compaction Criteria table of PP-28
(in AASHTO Provisional Standards[8]) did not
clearly indicate that they represent the pavement's cumulative ESALs for a 20-year
design life, rather than the cumulative
ESALs for a shorter or longer design life. The
WesTrack Forensics Team and the Lead
States Team both recently reminded users
that, regardless of the actual design life of
the pavement, the user should determine
the expected ESALs for 20 years and select
the design level for that traffic and loading.
For example, a project with a 5-year intended life may have a 5-year cumulative ESAL
count of 2.9 million. This corresponds to a

S U P E R PAV E M I X T U R E D E S I G N G U I D E

20-year cumulative ESAL count of 11.6 million (neglecting traffic growth compounding), and this latter ESAL count should be
used in the design. The 1999 version of PP28 includes a footnote to the Gyratory
Compaction Criteria table with the appropriate guidance.[8]
Experience has shown that rutting damage
often occurs in the first few years of a pavement's life; therefore, the design should be
based on the rate of loading. To properly
account for this in the mix design, the mix
designer should always use 20-year design
ESALs, essentially converting total loads to a
rate of loading. Estimating ESALs over a 20year life, instead of the actual design life,
may affect the mixture design compaction
level, the performance-graded (PG) binder
selection, and the aggregate consensus properties specified for the project. Compaction
criteria, aggregate properties, and volumetric properties are all more stringent at higher ESAL levels.

Superpave Performance-Graded
Binder Selection
The Superpave Performance-Graded Binder
Specification (AASHTO MP-1) is based on
providing a binder that is resistant to rutting,
fatigue cracking, and low-temperature cracking at specific pavement temperatures. The
binder temperature ranges in the specification are based on the high and low temperatures at which a binder reaches critical values of distress-predicting properties.
Reliability factors included in the design
method account for normal pavement temperature variations and allow the designer to
make a rational decision regarding the range
of temperature extremes for which to design.
Binder grade is selected based on design
high and low pavement temperatures
expected at the construction site and on
desired reliability.
The most common method of selecting a
binder grade is to determine the design air

temperature range for the specific project and

then to establish the corresponding design
pavement temperatures. Before selecting the
grade to be used, the designer must also consider traffic volume and traffic speed.
The owner should consider factors such as
cost and traffic levels in establishing reliability, and hence, the final binder grade selection, for a specific project. For example, if a
PG 64 binder provides 94 percent reliability
for high temperatures, it may not be costeffective to specify a PG 70 binder to obtain
98 percent reliability. However, if a PG 64
binder only provides 52 percent reliability, it
would probably be reasonable to specify a
PG 70 binder to obtain 98 percent reliability.
With respect to traffic volume, when the
design traffic exceeds 10 million ESALs,
Superpave suggests that an increase in the
high-temperature binder grade be considered. When design traffic is more than 30
million ESALs, Superpave requires a onegrade increase in the high-temperature
binder grade. With respect to traffic speed,
Superpave recommends increasing the hightemperature binder grade by one grade for
slow transient traffic (20 to 70 km/h) and by
two grades if standing traffic conditions
(<20 km/h) exist. The binder specifier
should increase the high-temperature grade
for traffic volume or traffic speed, but not for
both. If the system is used correctly, a pavement with high design ESALs with stopped
traffic conditions will require an asphalt
binder that is two high-temperature grades
higher than that required by the pavement
temperature alone.
It should be realized that when the hightemperature grade is increased by one grade,
the stiffness of the binder will approximately double. In other words, a PG 70 binder will
be twice as stiff as a PG 64 binder at a temperature of 64C. Furthermore, a PG 76
binder will be four times stiffer than a PG 64
binder at a temperature of 64C. Traffic
speed will also have an effect on binder stiffness in the pavement. At 50 km/h, a binder

2 S U P E R PAV E M I X T U R E D E S I G N G U I D E

will have a lower apparent stiffness than it

does when carrying traffic at 100 km/h. In
other words, a mixture containing PG 70
binder in a pavement with traffic moving at
50 km/h will have roughly the same stiffness as a mixture containing PG 64 binder in
a pavement with traffic moving at 100 km/h;
thus, the increased high-temperature grade
of the binder effectively offsets the effect of
slower traffic speeds.
Consideration should be given to the impact
that increasing the binder grade will have on
the construction process. Depending on the
grade, such an increase could require mixing
and compaction temperatures beyond reasonable construction temperatures.
Only strong aggregate skeletons can experience significant performance increases with
increased asphalt binder stiffness. The
stiffer binder locks the aggregate particles in
place to prevent rutting. The binder cannot
carry the load alone and cannot overcome a
poor aggregate skeleton by itself.
The final step before selecting the binder
grade to be specified is to compare the grade
being considered with grades historically
used in the area. If the binder seems unreasonably soft for preventing rutting based on
past history, or unreasonably stiff for construction purposes, the selected grade
should be reconsidered.

ity, flat and elongated particles, clay content, and gradation. Aggregate source properties, such as soundness, toughness, and
deleterious materials, were also found to be
important. However, the criteria applied to
the source properties were found to reflect
regional differences in aggregate quality,
and were usually based on aggregate availability. The panel determined that the
source properties were best left for each
State or local agency to establish. The following discussion addresses various aggregate properties (consensus and source) and
their effect on the Superpave design process.
Coarse Aggregate Angularity Mixtures
with crushed coarse aggregate with sharp,
angular shapes will usually have the greatest
shear resistance and, hence, the highest
resistance to rutting. These materials create
HMA mixtures with the highest voids in the
mineral aggregate (VMA). Coarse aggregate
angularity is defined as the percentage by
weight of the aggregate with one or more
fractured faces according to American
Society for Testing and Materials (ASTM)
D5821. Superpave requires increased percentages of crushed faces as the design ESAL
AG G R E GAT E P RO P E RT I E S

CONSENSUS PROPERTIES
(required)
coarse aggregate angularity (CAA)

Superpave Aggregate Selection

fine aggregate angularity (FAA)

Aggregates are the largest component of

HMA, making up 80 to 85 percent of the
mixture by volume and roughly 95 percent
of the mixture by weight. Aggregate characteristics and quality are major factors in the
performance of HMA. As part of its focus on
binder and mixture properties, in the early
1990's SHRP convened an expert panel to
determine which aggregate properties were
most important for pavement performance.
The properties selected included coarse
aggregate angularity, fine aggregate angular-

flat and elongated particles

clay content

SOURCE PROPERTIES
(agency option)
toughness
soundness
deleterious materials

S U P E R PAV E M I X T U R E D E S I G N G U I D E

C O N T R A S T I N G S TO N E S K E L E TO N S

limestones, have had less than 45 percent

(but more than 40 percent) uncompacted
void contents, but still have provided good
performance in pavements. If the performance has been satisfactory, the cubical manufactured fine aggregate may be used (with
caution).

Angular Aggregate

Rounded Aggregate

level increases. VMA increases somewhat as

coarse aggregate angularity increases.
Uncompacted Void Content of Fine
Aggregate (Fine Aggregate Angularity)
Similar to coarse aggregate, crushed angular
fine aggregate will usually have the greatest
shear resistance. The use of crushed angular
fine aggregate typically increases the mixture VMA. Fine aggregate angularity is established by AASHTO T304, Method A, which
measures the percentage of air voids present
in loosely compacted aggregate that passes
the 2.36-mm sieve. More fractured faces
generally result in higher uncompacted void
contents in this test. Superpave specifies
uncompacted void contents of at least 45
percent on high-volume roads (>3 x 106
ESALs). Crushed manufactured fine aggregates generally have uncompacted void contents of at least 44.5 percent, while rounded
natural sands typically are less than that
value. When a fine aggregate known to be
angular has test results lower than expected,
the aggregate's bulk specific gravity should
be verified since the test result is sensitive to
this property; a significant change in the
bulk specific gravity should trigger a
redesign of the mixture.
Particle shape can also influence the uncompacted void content. Some very cubical manufactured fine aggregates, especially some

The fine aggregate's uncompacted void content significantly influences the VMA. The
use of cubical angular fine aggregate is recommended to increase the VMA. Care
should be taken when using aggregate with
uncompacted void contents higher than 47
percent; use of these aggregates may result
in mixtures with excess VMA, which leads,
in turn, to a very high binder content.
Flat and Elongated Particles The percentage of flat and elongated particles (not
flat or elongated) in coarse aggregate is
another important aggregate parameter. Flat
and elongated particles can break during the
construction process, changing the mixture
gradation and the overall mixture properties.
Soft aggregate has a greater tendency to
break than hard aggregate. Flat, slivered
aggregate particles also have a tendency to
lie flat in the pavement, creating slippage
planes and reducing aggregate interlock. A
small percentage of flat and elongated particles in the mixture may increase the VMA in
the laboratory-designed mix. A further
increase may, however, decrease the VMA in
the plant-produced mixture because of
aggregate breakage during mixing.
The critical measurement for a flat and elongated particle is the ratio of its maximum
and minimum dimensions. Current
Superpave standards allow no more than 10
percent of the coarse aggregate particles to
be flat and elongated (i.e., a ratio greater
than 5:1). Testing is performed according to
ASTM D4791, "Flat Particles, Elongated
Particles, or Flat and Elongated Particles in
Coarse Aggregate." Superpave establishes
that testing be done on material retained on
the 4.75-mm sieve, instead of on the 9.5-mm
sieve as specified in the ASTM method.

4 S U P E R PAV E M I X T U R E D E S I G N G U I D E

Testing aggregate particles passing the 9.5mm sieve and retained on the 4.75-mm sieve
will be more difficult and results may be
more variable.
Sand Equivalent
Sand equivalent, as
measured by AASHTO T176, "Plastic Fines
in Graded Aggregates and Soils by Use of the
Sand Equivalent Test," identifies the presence of clay in the fine aggregate. Clay can
make the mixture moisture sensitive and/or
combine with moisture to cause the mixture
to act "tender" (i.e., to lose density with continued compaction in the field). Clay content must be controlled by satisfying the
minimum sand equivalents specified in the
Superpave standards.
Aggregate Toughness Typically, mixtures
containing very hard aggregate (i.e., a Mohs
hardness of 7 or greater) do not have a problem meeting VMA criteria. A very hard
aggregate, such as basalt, does not easily
crush or degrade during laboratory compaction or during mix production in an HMA
plant. These aggregates can produce mixtures that have an adequate VMA.
Soft aggregates, such as some types of limestone having a Mohs hardness of about 5, are
often abraded during the gyratory compaction process; this can make it difficult to
meet VMA criteria during the design phase.
During production, the aggregates are often
abraded in the hot-mix plant to an even
greater degree than in the laboratory design
using a gyratory compactor. When plantproduced material is compacted in a gyratory compactor, the aggregate is abraded further and even more fines are generated in
the mixture; this further reduces the VMA.
Mixtures designed with soft aggregates often
have a problem meeting VMA criteria in the
design stage and, particularly, during production. It is extremely important that the
plant-produced mixture satisfy the minimum VMA requirement.
Mixtures designed with a blend of hard and
soft aggregate could have difficulty meeting

VMA specifications. The addition of hard,

coarse, or fine aggregate to these types of
aggregate blends will usually increase the
VMA.

Superpave Mixture Design

Considerations
Superpave mixture design criteria include
air voids, VMA, and voids filled with asphalt
(VFA). Meeting the VMA minimum criterion
is usually difficult to achieve during mix
design and typically even more difficult to
achieve in the plant-produced material. This
document will only discuss VMA.
Voids in the Mineral Aggregate In many
cases, achieving minimum VMA requirements during the design phase can be difficult. Many factors affect VMA. The most critical of these are aggregate characteristics
such as gradation, surface texture, and
shape. If the design VMA is close to (i.e., no
more than 0.6 percent above) the minimum,
aggregate properties may change during production and cause the VMA to drop below
the minimum during mixture production.
Differences between as-designed and plantproduced properties and other field problems are documented in the NAPA publication, Field Management of Hot-Mix Asphalt.[9]
As noted above, VMA is affected by both the
aggregate gradation (relationship to the
aggregate maximum density line) and the
aggregate's characteristics and properties.
For all designs, VMA should be plotted as a
function of binder content and the resulting
graph should be evaluated to check the VMA.
Typically, VMA will decrease with increasing
binder content to some minimum, then
increase as binder content continues to
increase. The design binder content, selected
at 4 percent air voids, should be near the
minimum of the plotted curve or preferably
on the lean binder content side of the curve.
If the VMA at the design binder content is on
the rich side of the VMA curve, adjustments
to the gradation should be considered; these

S U P E R PAV E M I X T U R E D E S I G N G U I D E

are discussed later in this section. In

Superpave mixtures, however, VMA is sometimes insensitive to binder content and
shows little change. If the VMA at the design
binder content is close to the minimum
allowable VMA value and the curve is relatively flat, the mixture should be redesigned.
There are two competing demands during
the mix design process: sufficient inter-particle space must be available for a minimum
amount of binder, but, at the same time, the
aggregate must have a sufficiently strong
skeleton to carry the traffic loads. Superpave
mixture design specifications require that
adequate VMA be obtained without weakening the aggregate skeleton.
Having representative aggregate bulk specific gravity values is necessary in order to
accurately calculate a mixture's VMA during
design and production. For this reason,
aggregate bulk specific gravity should be
determined at a frequency appropriate for
the variability of the source.
Mixtures with high VMA need to be
reviewed for possible performance problems. The WesTrack Forensic Team recommended that the VMA of coarse-graded
Superpave mixtures be no more than 2.0
percent above the minimum required
value.[10] Furthermore, the Team recommended running a draindown test (AASHTO
T305-97) on these mixtures if the VMA is 1.5
percent or more above the minimum value.
If a gradation yields a mixture with too high
of a VMA and, consequently, too high of a
binder content, the mixture design should
be repeated with a new gradation with lower
VMA.
Gradation Effect Problem mixes typically will have a low VMA and may not be
responsive to changes in gradation (when
aggregate sources are not changed). Usually,
however, changing the gradation of a mixture
will influence the amount of void space in the
aggregate skeleton. The effect of gradation is
separate from the shape and surface texture

effects if all particles have the same shape

and texture. If the stockpiles in the blend are
of dissimilar materials, changing the stockpile
percentages will change the gradation, but it
will also influence the shape and texture of
the aggregate blend. Thus, VMA will change
not only because of gradation changes, but
also because of shape and texture changes.
Research papers published by Nijboer in the
1940's, Goode and Lufsey in the 1960's, and
the Asphalt Institute in the 1980's provide a
basis for the 0.45 power chart. Nijboer investigated aggregate gradations plotted as the
log percent passing versus log particle size.
He showed a maximum packing density for
both gravel and crushed aggregates when
the slope was 0.45. Goode and Lufsey confirmed Nijboer's results on gravel aggregates.
Work by the Asphalt Institute evaluated the
maximum density line on a 0.45 power chart
for both gravel and crushed limestone mixtures and reconfirmed the previous results.
Moving the gradation away from the 0.45
power maximum density line generally
increases the VMA for a fine gradation, i.e.,
when the gradation is above the maximum
density line. For a coarse gradation, VMA
may decrease slightly and then increase as
the gradation moves away from the maximum density line. For hard aggregates, the
Job Mix Formula (JMF) should be parallel to
the maximum density line until after passing the restricted zone, i.e., for aggregate
retained on the 4.75-mm sieve (for 25-mm or
larger mixes) or on the 2.36-mm sieve (for
19-mm or smaller mixes). Then the gradation line should be taken to the desired 0.075-mm content. Therefore, with fine gradations, the JMF should be above and parallel to the maximum density line. For a
coarse gradation, it should be below and parallel to the maximum density line.
Many coarse Superpave mixes have an "S"
shape, starting on the fine side of the maximum density line and finishing the S on the
coarse side. If the same particle shape and
texture are used (same aggregate source, dif-

6 S U P E R PAV E M I X T U R E D E S I G N G U I D E

If the aggregate is obtained from a quarried

source, the crushed material should not be
placed into a single stockpile, but should be
divided into at least three separate size
ranges, depending on the nominal maximum size of the aggregate required in the
mix. The use of multiple stockpiles allows
more flexibility to change gradation and,
thus, VMA. In addition, it may still be necessary to incorporate another size of aggregate
from the quarry or from a different source.
The VMA of coarse-graded mixes can generally be increased by reducing the amount of
material passing the 4.75- and 2.36-mm
sieves. The reason has to do with packing
smaller particles fill the spaces between larger ones. By reducing the amount of material
passing the 4.75- and 2.36-mm sieves, intermediate material is removed and more space
is created between the coarse aggregate particles. Hence, the mixture cannot compact
as tightly, i.e., VMA is increased.
In fine-graded mixes, VMA is created by fine
aggregatethe material passing the 2.36-mm
sieve. To increase the VMA in fine-graded
mixes, the percentage of material passing
the 2.36-mm sieve should be increased. Care
should be taken not to create a hump in the

The dust content (i.e., the amount of material finer than 0.075 mm) in a mixture has a
significant effect on the VMA. Lowering the
dust content will increase the VMA. This
effect may not be entirely due to the gradation, but may also be due to characteristics
of the dust, such as shape and size. In general, reducing dust content to the extent that
the dust-to-binder ratio will allow will maximize the amount of VMA that can be
obtained for the specific gradation.
If the dust content is coming from the addition of mineral filler, adjusting the dust conF H WA 0 . 4 5 P OW E R G R A D I N G C H A RT
100
Maximum
Size

80

Percent Passing

If the aggregate is obtained from a gravel

source, normally the fine aggregate must be
removed before the coarse aggregate enters
the crusher. The crushed material should be
divided into three or more stockpiles that
can then be blended into a combination that
meets the minimum VMA requirements. All
of the aggregate processed may not be
usable; it may be necessary to waste some of
the material in order to meet the requirements of the mixture design.

gradation on the 0.6- to 0.3-mm sieves using

an aggregate that has a low fine-aggregate
angularity value.

60
40
20
0

Maximum Density Line

.075 .3 .6

1.18

2.36

4.75

9.5

12.5

19.0

Sieve Size (mm) Raised to 0.45 Power

S U P E R PAV E AG G R E GAT E G R A DAT I O N

100

Percent Passing

ferent sizes), the highest VMA that can be

achieved is that for the gradation that is the
farthest from the maximum density line. To
minimize the chance of mix tenderness during construction, the JMF should be 3 to 4
percent above the lower control point for the
2.36-mm sieve.

Design Aggregate Structure

0
.075

.3

2.36

12.5

19.0

Sieve Size (mm) Raised to 0.45 Power

S U P E R PAV E M I X T U R E D E S I G N G U I D E

tent can be simply a matter of reducing the

amount of filler being used. If the dust is predominantly from one of the aggregate stockpiles, e.g., screenings, reducing the percentage of that stockpile used in the blend should
be tried. If the screenings are the only manufactured fines coming into the mix, using
washed screenings or blending with a
washed screening may be necessary.
If baghouse fines will be introduced back into
the mix during production, some of the same
fines should be added during the mix design.
During the design, adding half of the quantity of baghouse fines expected to be added
during production is an appropriate procedure. These fines should be obtained from
the actual plant that will be used for production; otherwise, mineral filler or an alternate
source of baghouse fines could be used.
These fines will reduce the VMA of the mixture. If the aggregate in the mix contains friable particles, a greater quantity of dust
should be used in the laboratory mix design
since the friable particles tend to create more
dust during mix production. A mix design
that includes baghouse fines will be more
representative of the mix as produced. The
addition of baghouse fines during the mix
design will better simulate the reduction in
VMA that typically occurs during production.
Surface Texture Effect The way in which
aggregate particles pack together for any
given gradation is influenced by the surface
texture of the particles. Rougher texture generates more friction between aggregate particles and the mixture therefore resists compaction. Hence, for a given number of
design gyrations, the mixture will not compact as much and the VMA will be higher.
Smooth texture, by contrast, does not generate as much friction between aggregate particles. For a given number of design gyrations, the mixture containing smoother particles will compact more easily and the VMA
will be lower.
Typically, crushed faces have more texture
than uncrushed faces. In the case of gravel

aggregate, the uncrushed portion of the

particles tends to have a smooth texture.
The greater the percentage of each individual particle surface area that is fractured,
the more surface texture that will be present. Usually, the more a gravel is crushed,
the more surface texture it will have.
Particles with two crushed faces tend to
have a greater percentage of surface area
with rough texture than will particles with
only one crushed face. However, crushing
will not always increase texture, because
some aggregates fracture with very smooth
faces.
If manufactured sand and natural sand are
being used together in a mix design, the percentage of manufactured sand can be
increased to increase surface texture.
Substituting 20 percent washed manufactured sand (with good "bite") for an equivalent amount of natural sand can increase the
VMA substantially. (What is good bite?
Squeeze a handful of angular manufactured
sand, then a handful of rounded natural
sand, and feel the difference in the way the
particles bite into one another.) If the manufactured sand contains more dust than the
natural sand, gains in VMA from the surface
texture may be decreased by the increase in
dust content. For example, if the natural
sand is relatively clean and the manufactured sand has a high minus 0.075-mm dust
content, the benefit of increased surface texture may be partially or completely offset by
the increased dust content.
Shape Effect For any given gradation,
the density to which aggregate particles will
pack is influenced by the shape of the particles. Angular particles (i.e., those with
sharp, defined edges) tend to produce mixtures with a higher VMA than mixtures containing rounded particles. Cubical particles
that retain a sharp, angular edge tend to create a higher VMA than particles with rounded edges.
The effect of flat and elongated particles
depends on the laboratory compaction

8 S U P E R PAV E M I X T U R E D E S I G N G U I D E

method. Under Marshall compaction, the

particles were not as free to rotate as they
are in a Superpave gyratory compactor. In
fact, flat particles tend to bridge in a
Marshall mold and give a high VMA.
Therefore, flat and elongated particles tend
to increase asphalt content in a Marshall mix
design. In the Superpave gyratory compactor, where the particles are kneaded into
a more stable condition, flat and elongated
particles tend to lie horizontally; this
reduces the VMA and the optimum binder
content.
During construction, rollers tend to orient
flat and elongated particles horizontally. A
Marshall mix design containing excess flat
and elongated particles could compact very
easily or be compacted to a lower air void
content than desired during the roadway
compaction process. A Superpave mix
design will have a more appropriate binder
content since gyratory compaction better
simulates compaction during construction
than does Marshall compaction. Therefore,
the influence of particle shape must be considered when comparing the VMA of
Marshall specimens to that of Superpave
specimens.
If a mix design has a low VMA, the amount of
flat and elongated particles must be determined. Superpave specifications limit the
percentage of particles with a maximum-tominimum dimension ratio of greater than 5.
If flat and elongated particles are contributing
to a low VMA in a mixture, the percentage of
particles that exceed a 3:1 ratio should be
determined. If the percentage of particles
exceeding the 3:1 ratio is high (i.e., greater
than 40 percent), material from a coarse
aggregate stockpile that has a lower percentage should be added. It may be possible to
change one of the coarse aggregate stockpiles
for another that contains more cubical and
angular aggregate particles. Adding an intermediate-size coarse aggregate with cubical
and angular shapes will prevent the larger
particles from lying flat. Thus, VMA will
increase.

Crushing operations influence the amount

of flat and elongated particles produced. If
excess flat and elongated particles are being
produced, the crushing operation should be
evaluated. In some instances, the amount of
flat and elongated particles produced can be
reduced by changing the aggregate feed rate,
or by changing the opening of the cone or
jaw crushers. In some cases, it might be necessary to modify the crushing operation by
adding to or changing the equipment used.
Vertical-shaft impact crushers, for example,
tend to produce more cubical particles than
do some cone crushers (especially older
models).
In summary, VMA depends on the gradation, surface texture, and particle shape of
the aggregate. In designing a mix, all of these
characteristics must be considered. When
there is difficulty in meeting the minimum
VMA requirements, some or all of the above
characteristics must be adjusted. It should be
remembered that the VMA of a plant-produced mixture is typically lower than the
VMA of the laboratory trial mix formula.
Allowances should be made for the reduction in VMA that will occur between the laboratory-designed and the plant-produced
mixtures.[9]
Mixing and Compaction Temperatures
For unmodified binders, the mixing and
compaction temperatures used during the
design process should be established with a
rotational viscometer. If the binder is modified, the binder supplier must provide recommended mixing and compaction temperatures. If the binder content determined in
the mix design process seems unrealistic,
the supplier should be consulted to determine whether the mixing and compaction
temperatures being used are still appropriate
for the material being delivered. The compaction temperature used during design
should also be used in plant production quality control and quality assurance testing.
The laboratory mixing and compaction temperatures may not be appropriate for use in

S U P E R PAV E M I X T U R E D E S I G N G U I D E

actual plant production and laydown.

Environmental conditions at the time of construction and other factors, such as haul
length and lift thickness, need to be considered in establishing the actual mixing temperature used at the plant. These temperatures should not be any greater than those
necessary to ensure complete mixing of the
HMA while minimizing premature aging of
the binder and providing for adequate compaction in the field.
If the quality control and/or quality assurance sample is taken behind the paver and
the sample requires reheating before compaction, a comparison should be made
between the properties of the reheated samples and samples that are compacted before
cooling. Detailed procedures for quality
assurance sampling and testing should be
established before construction begins.
Dust-to-Binder Ratio
Superpave calculates the dust-to-binder ratio using the effective binder content. Using the effective
binder content rather than the total binder
content will normally result in a higher dustto-binder ratio because of binder absorption
into the aggregate. To account for absorption, the limit for the dust-to-binder ratio
should be increased. In the original
Superpave specification, the dust-to-binder
ratio was 0.6 to 1.2 by weight. FHWA's
Asphalt Mixture Expert Task Group and the
AASHTO Lead States recommended changing the limit to 0.6 to 1.6; AASHTO subsequently added a note to the mix design specification suggesting that agencies consider
changing the limit for coarse-graded mixes to
0.8 to 1.6.

amount that can be added to a mixture.

Changing the fines source or production
process will change how the fines affect the
mixture characteristics. For fine-graded mixtures (above the maximum density line at the
critical sieve), a ratio of about 1.0 has provided satisfactory performance. For coarse-graded mixtures (below the maximum density
line at the 4.75-mm sieve), as the VMA
increases the dust-to-binder ratio should
increase toward 1.6. If the mixture VMA is
high (more than 2.0 percent above the minimum), the ratio should approach 1.6.
High dust-to-binder ratios will typically stiffen the mixture and improve permanent
deformation resistance. However, if the
VMA is more than 1.5 percent above the
minimum, it is preferable to adjust the
aggregate properties to reduce the VMA
instead of increasing the dust content.
Performance Indicator Tests No test is
currently available that is satisfactory, by
itself, as a performance predictor for mixtures generated by Superpave volumetric
procedures. Appendix A contains a discussion of various tests that may be used to indicate the relative performance of different
mixtures. The designer should have experience with one of the tests before assuming
that the test results will actually predict field
performance. Criteria developed elsewhere
may not apply to a particular combination of
materials, environmental conditions, pavement structure, and traffic.
Completing the Design After the Superpave design is completed, the designer needs
to ask two final questions:

During design, mixtures that are above the

maximum density line at the 2.36-mm (for
19-mm or smaller mixtures) or 4.75-mm (for
25-mm or larger mixtures) critical sieve
should have a dust-to-binder ratio of no more
than 1.4. For mixtures that pass below the
maximum density line at the critical sieve
size, the ratio should not exceed 1.6.
Characteristics of the fines will control the

1 0 S U P E R PAV E M I X T U R E D E S I G N G U I D E

Is this HMA design reasonable

and logical?
Is the binder content reasonable
for the type of aggregate, the nominal maximum aggregate size, the
VMA, and the gradation used in
the mixture?

If the answer to either question is "No," the

design should be re-evaluated and/or
redone. Once the answers to both questions
are "Yes" and mixing plant operation has
begun, the mixture volumetrics of the plantproduced mixture must be checked and
those volumetrics must meet the minimums
required in the design. In addition, performance tests should be repeated on the plantproduced mixture if these tests were per-

formed during the original design phase.

During plant verification, enough mixture
should be produced to ensure that the plant
is operating uniformly. The designer should
be prepared to make mixture adjustments to
account for changes caused by plant production. If changes are made, the mixture
should be re-verified. A key to good mixture
performance is to verify the HMA plant-produced mixture properties.

DESIGN CHECK LIST

Use a performance-graded (PG) binder and an N-design value

appropriate for the weather, traffic level, and traffic speed for the
project under consideration. Heavy, slow traffic will require a
stiffer PG binder than may have been used in the past.

Check that a complete mix design has been done in accordance

with specifications and that it meets all of the aggregate consensus property requirements and specified volumetric criteria.

Check that the submitted design contains a reasonable binder

content for the materials used and the design level specified.

Generally, more dust (material passing the 0.075-mm sieve) is

needed for coarse-graded mixtures. The character of the dust
will control how much can be added to the mixture. Laboratory
samples should contain the expected plant-produced amount of
material finer than 0.075 mm.

In coarse-graded mixtures, if the VMA is more than 1.5 percent

above the specified minimum, check for binder draindown.
Excessive draindown is an indication that the binder content is
too high for the binder grade, aggregate type, and/or gradation
being used.

Evaluate the mixture with a performance indicator test that has

worked satisfactorily based on local experience (until a universally acceptable test is included in Superpave). Does the mixture
perform as expected?

Verify the properties of the plant-produced mixture to check volumetric properties. Repeat the performance test on the plantproduced mixture if the test was run during the mixture design.

S U P E R PAV E M I X T U R E D E S I G N G U I D E

11

PERFORMANCE TESTS

Introduction
Many different types of performance tests
are currently available for assessing a mixture's ability to resist permanent deformation (commonly referred to as "rutting").
These tests, which include Marshall
flow/stability, Hveem stability, the gyratory
testing machine, wheel-track testers, the
Superpave Shear Test device, and triaxial
testers, generally attempt to quantify mixture strength and/or stiffness. The individual tests have shown varying levels of success in capturing a mixture's ability to resist
rutting. Therefore, the designer must know
the limitations of each test and how to incorporate test results into mixture design selection. This appendix describes each test and
examines how suitable each is for assisting
engineers in designing rut-resistant mixtures. At the same time, mixture designers
are reminded that a mixture that is resistant
to rutting will not necessarily resist thermal
or fatigue cracking, moisture damage, or
durability problems such as raveling.

Marshall
The Marshall mixture design process seeks
to optimize a mixture's performance with
regard to fatigue cracking, rutting, and durability by determining the optimum binder
content for the gradation selected. Once the
optimum binder content is selected, the
mixture must meet minimum stability values and maximum flow values. A number of
European countries have modified the specification criteria to use a stability quotient
(stability/flow) criterion in lieu of the minimum stability and maximum flow values.

Many mixtures have stability values that are

two or three times the minimum, but also
exceed the maximum flow value. The
European approach appears more logical
because it normalizes the stability/flow values. Marshall flow does provide an indication when a mixture is over-asphaltedhigh
flow values indicate excess binder content.
The Marshall test conditions may significantly affect the test's value in predicting
rutting performance. First among these is
the ratio of the test specimen's size to the
nominal maximum aggregate size. A 100mm- (4-in.-) diameter specimen that
includes a large nominal maximum aggregate size (37.5 mm) or a more open-graded
mixture (one containing little intermediatesize material) does not provide good-quality
test data. The effects of the specimen edges
are amplified and the assumption that the
Marshall breaking head is applying a uniform load across the specimen is no longer
valid. The effective load on the specimen
(load divided by the contact area) is higher
for larger nominal maximum aggregate size
mixtures. Another shortcoming of the procedure is the 60C (140F) temperature at
which the Marshall test is conducted. The
mixture may encounter temperatures 5 to
10C (9 to 18F) higher in place in some
parts of the country.

Hveem
The Hveem Stabilometer is a mixture design
tool used primarily in the western United
States. The concept behind the Hveem
Stabilometer is an empirical measurement
of the internal friction within a mixture,

1 2 S U P E R PAV E M I X T U R E D E S I G N G U I D E

resulting from application of a vertical axial

load. Like the Marshall Method, Hveem testing is conducted on 100-mm- (4-in.-) diameter specimens at 60C (140F). As noted
above, this temperature does not always represent the highest temperature a mixture
will experience in the field. Furthermore,
stabilometer values are measurements of
internal friction, which is more a reflection
of the properties of the aggregate than of the
binder. As with Marshall flow values, Hveem
stability does provide an indication when a
mixture is over-asphaltedlow stability values indicate excess binder content.

Gyratory Testing Machine

The gyratory testing machine (GTM), developed by the U.S. Army Corps of Engineers,
measures the increase in the angle of gyration during compaction. The gyratory shear
index, a measure of a mixture's stability, is
the initial angle of gyration divided by the
maximum angle. Shear indices above 1.1
usually indicate poor mixture stability, while
values nearer to 1.0 are more stable.

Wheel-Track Testers
Currently, three wheel-track testers are
available commerciallythe French LCPC
[Laboratoire Central des Ponts et Chaussees]
Rutting Tester, the Georgia Loaded-Wheel
Tester (marketed as the Asphalt Pavement
Analyzer), and the Hamburg Wheel-Tracking
Device. Conceptually, the three devices are
the same (a rolling load is applied to laboratory-scale specimens), but they differ signif-

icantly in design, load configuration, and

test conditions. To complicate the comparison, each device has a different recommended pass/fail criterion for mixtures. The
machine design for each of the devices significantly affects how well its results can be
correlated with field performance.[11]
The French LCPC Rutting Tester uses a 90mm-wide pneumatic tire to test specimens
that are 180-mm wide. This specimen width
and the closeness of the confining rigid specimen holder to the location of repeated loading distorts the development of the mixture's
shear plane, especially for mixtures containing larger aggregate. As a result, poor mixtures tend to perform better than expected
in the French device, and discriminating
between good- and poor-performing mixtures becomes difficult. The device should
not be used to test mixtures that have aggregate larger than 20-mm.
The Georgia Loaded-Wheel Tester (GLWT)
runs a concave steel wheel over a pressurized 29-mm-wide hose to apply loads on
specimens. Testing can be conducted on dry
specimens or underwater. For mixtures containing a larger size of aggregate, aggregate
bridging becomes a problem. The applied
footprint from the pressurized hose is much
narrower than the footprint of a vehicle tire
that the mixture will be subjected to under
field conditions. As a result, the GLWT test
criteria may allow for some poor mixtures to
be placed.
The Hamburg Wheel-Tracking Device
(HWTD) applies a sinusoidal load on specimens using a steel wheel underwater at an
elevated temperature. The HWTD measures
a mixture's ability to resist rutting and strip-

S U P E R PAV E M I X T U R E D E S I G N G U I D E

13

ping. The probability that these same test

conditions will coincide in the field is unlikely. The use of a steel wheel further increases the severity of the test. Because a steel
wheel does not deform under the test conditions like a pneumatic tire, the effective load
per unit area is much higher than that occurring during actual field loading. A mixture
that survives the HWTD test should be rutresistant in the field; however, mixtures that
do not survive the test may also perform
well in the field. Use of this device in mixture pass/fail situations can result in the
rejection of acceptable mixtures.
FHWA's Asphalt Mixture Expert Task Group
recommends the following cautionary note
for wheel-track testers:
Rut testers, properly calibrated,
have been utilized by some agencies
as effective proof testers. However,
they should not be used to predict
actual pavement performance
because of differences in in-service
temperature and loading conditions.
The devices use empirical evaluation of some measured response to a
loaded wheel as an indicator of performance. Local criteria from one
region are not applicable in another.
As such, each potential user needs
to develop his/her own evaluation
of wheel test results using local conditions.

and frequency sweep at constant height. All

but the repeated shear at constant height
test were included in the original Superpave
performance testing program. The report,
Background of SUPERPAVE Asphalt Mixture
Design and Analysis,[4] describes the test
modes in detail. Problems have been
encountered in interpreting data from the
repeated shear at constant stress ratio test,
the simple shear at constant height test, and
the frequency sweep at constant height
tests.[12-14] As a result of these problems, no
attempt was made to link the predicted performance from the laboratory tests to the
field performance.
Romero and Mogawer presented additional
SST results and compared the results of
repeated shear at constant height tests with
those from full-scale accelerated tests.[15]
They stated that the repeated shear at constant height test mode is able to rank mixtures with different binders, but with high
variability in mixture stiffness. This variability often makes it impossible to place each
mixture into statistically different groups.
SST results have shown significant variability between laboratories for the simple shear
at constant height test mode. Until this variability can be reduced, it will not be possible
to adopt universally acceptable criteria. In
summary, the SST is still being studied to
determine the usefulness of the results from
each of its six test modes; work with the
device has not reached a point where its
results can be used in any standard mode to
predict rutting performance.

Superpave Shear Tester

The Superpave Shear Tester (SST) can be
operated in any of six different modes: volumetric, uniaxial strain, repeated shear at
constant stress ratio, repeated shear at constant height, simple shear at constant height,

Creep Tests
Triaxial testing equipment has been used for
many years in soil mechanics and on
asphalt materials. The creep test and, to a

1 4 S U P E R PAV E M I X T U R E D E S I G N G U I D E

lesser extent, the creep-creep recovery

(CCR) test have been used for HMA under
various triaxial stress states. The creep and
CCR tests are used to estimate rutting
potential. Most commonly, a uniaxial static
test is used in either a confined or an
unconfined mode. The unconfined test does
not simulate field conditions. The applied
pressure cannot exceed 207 kPa (30 psi)
without specimens failing, and the test temperature is kept at 40C (104F) in the
unconfined test, well below actual field
loading conditions that often reach 830 kPa
(120 psi) and 60C (140F). The confined
creep test can be run at higher pressures
and temperatures, with a confining pressure
of 138 kPa (20 psi). Research has shown that
confined creep testing has a higher correlation to permanent deformation than unconfined testing.[16] A viscoelastic layered pavement performance system that uses creep
and CCR testing to estimate the permanent
deformation in asphalt mixtures subjected
to repeated haversine loading of in-service
loading frequencies already exists.[17]
However, this CCR testing does not exist as
input directly into constitutive models for
asphalt pavements. Research is underway to
examine the ability of this equipment to
measure "fundamental" material properties
and to include these measurements in constitutive modeling. Currently, the equipment and procedures to help engineers
make rational mixture design decisions are
not available in the context of measuring
engineering properties as input to constitutive models.
One test that is being recommended for performance evaluation of HMA by the
researchers on NCHRP Project 9-19 is the
Static Creep/Flow-Time test. In this test, a
cylindrical sample of bituminous paving
mixture is subjected to a static axial load.
The test can be performed without confine-

ment or with a confining pressure applied to

better simulate in situ stress conditions. The
flow time is defined as the time after initial
load application when shear deformation,
under constant volume, starts. The applied
stress and the resulting permanent and/or
axial strain response of the specimen are
measured and used to calculate the flow
time. Using this test, the selection of the
design binder content and aggregate structure can be fundamentally enhanced by the
evaluation of the mix's resistance to shear
flow (flow time). This fundamental engineering property can be used as a performance criteria indicator for permanent deformation resistance of the asphalt concrete
mixture, or can simply be used to compare
the shear resistance properties of various
bituminous paving mixtures.

Conclusions
Currently, no single test is suitable as a
national standard for predicting rutting. The
development of such a procedure is urgently
needed, but a satisfactory procedure may be
years away. In the meantime, if an agency
has extensive experience with a particular
test over a range of materials typical of its

Currently, no single test is suitable as a

national standard for predicting rutting.
geographic area, it should consider using the
test to predict rutting performance. Each of
the devices outlined here has difficulty in
predicting the true performance of an
asphalt mixture and should be used with
great caution.

S U P E R PAV E M I X T U R E D E S I G N G U I D E

15

1. Superpave 1999-2000 National

Implementation, Report to the
AASHTO Task Force on SHRP
Implementation, May 2000.

2. Superpave Construction
Guidelines, Special Report 180,
National Asphalt Pavement
Association, Lanham,
Maryland, 1997.

3. McGennis, R.B.; Shuler, S.; and

Bahia, H.U. Background of
SUPERPAVE Asphalt Binder Test
Methods, Report No. FHWA-SA94-069, Federal Highway
Administration, Washington,
D.C., 1994.

4. McGennis, R.B.; Anderson,

R.M.; Kennedy, T.W.; and
Solaimanian, M. Background of
SUPERPAVE Asphalt Mixture
Design and Analysis, Report No.
FHWA-SA-95-003, Federal
Highway Administration,
Washington, D.C., 1995.

5. The SUPERPAVE Mix Design

System Manual of Specifications,
Test Methods, and Practices,
Report No. SHRP-A-379,
National Research Council
(Strategic Highway Research
Program), Washington, D.C.,
1994.

1 6 S U P E R PAV E M I X T U R E D E S I G N G U I D E

6. Superpave Performance-Graded
Asphalt Binder Specification and
Testing, Superpave Series No. 1
(SP-1), The Asphalt Institute,
Lexington, Kentucky, 1995.

7. Superpave Mix Design,

Superpave Series No. 2 (SP-2),
The Asphalt Institute,
Lexington, Kentucky, 1996.

8. AASHTO Provisional Standards,

Interim Edition, American
Association of State Highway
and Transportation Officials,
Washington, D.C., May 1999.

9. Field Management of Hot Mix

Asphalt, Report No. IS-124,
National Asphalt Pavement
Association, Lanham,
Maryland, 1997.

10. WesTrack Forensic Team.

Performance of Coarse-Graded
Mixes at WesTrackPremature
Rutting, Report No. FHWA-RD99-134, Federal Highway
Administration, Washington,
D.C., June 1998.

11. Williams, R.C., and Stuart, K.D.

"Evaluation of Laboratory
Accelerated Wheel Test
Devices," Proceedings, The 9th
Road Engineering Association
of Asia and Australasia
Conference, Wellington, New
Zealand, May 1998.

12. Hicks, R.G., and Finn, F.N. Stage

1 Validation of the Relationship
Between Asphalt Properties and
Asphalt-Aggregate Mix
Performance, Report No. SHRPA-398, National Research
Council (Strategic Highway
Research Program), Washington,
D.C., 1994.

13. Zhang, X. "Evaluating

Superpave Performance
Prediction Models Using a
Controlled Laboratory
Experiment," Journal of the
Association of Asphalt Paving
Technologists, Volume 66, 1997.

15. Romero, P., and Mogawer, W.S.

"Evaluation of the Superpave
Shear Tester Using 19-mm
Mixtures From the Federal
Highway Administration's
Accelerated Load Facility,"
Journal of the Association of
Asphalt Paving Technologists,
Volume 67, 1998, pp. 573-601.

16. Roberts, F.L.; Kandhal, P.S.;

Brown, E.R.; Lee, D.Y.; and
Kennedy, T.W. Hot Mix Asphalt
Materials, Mixture Design, and
Construction, National Asphalt
Pavement Association
Education Foundation, Lanham,
Maryland, 1991.

17. VESYS 3am User's Manual,

Office of Research and
Development, Federal Highway
Administration, McLean,
Virginia, 1996.

14. Romero, P., and Mogawer, W.S.

"Evaluation of the Superpave
Shear Tester's Ability to
Differentiate Between Mixtures
With Different Aggregate Size,"
Transportation Research Record
1630, National Research
Council, Washington, D.C.,
1998, pp. 69-76.

S U P E R PAV E M I X T U R E D E S I G N G U I D E

17

Team Members
Ray Brown, Director
National Center for Asphalt Technology
211 Ramsay Hall
Auburn University, AL 36849
Tel.: (334) 844-6228
Fax: (334) 844-6248

Gerald Huber
Heritage Research Group
7901 W. Morris Street
Indianapolis, IN 46231
Tel.: (317) 243-0811
Fax: (317) 486-5095

Larry Michael, Regional Engineer

Maryland Department of Transportation
528 East Main Street
Hancock, MD 21750
Tel.: (301) 678-6134
Fax: (301) 678-5190

Ron Sines, Director, QC/QA Operations

P.J. Keating Co.
P.O. Box 367
Fitchburg, MA 01420
Tel.: (978) 582-5200
Fax: (978) 582-7130

Erv Dukatz, V.P., Materials & Research

Mathy Construction Company
P.O. Box 563
915 Commercial Court
Onalaska, WI 54650
Tel.: (608) 779-6392
Fax: (608) 781-4694

Jim Scherocman, Consulting Engineer

11205 Brookbridge Drive
Cincinnati, OH 45249
Tel.: (513) 489-3338
Fax: (513) 489-3349

Liaison Members
John D'Angelo, Sr. Pavement Mtls. Engineer
Office of Pavement Technology (HIPT-10)
Federal Highway Administration
Room 3118, Nassif Building
Washington, D.C. 20590
Tel.: (202) 366-0121
Fax: (202) 366-7909
E-mail: john.d'angelo@fhwa.dot.gov

Chris Williams (formerly Research Highway

Engineer, Federal Highway Administration),
Asst. Prof., Dept. of Civil & Env. Eng.
Michigan Technological University
870 Dow Env. Sci. & Eng. Building
Houghton, MI 49931
Tel.: (906) 487-1630
Fax: (906) 487-2943

Editor
Terry Mitchell, Research Mtls. Engineer
Office of Infrastructure Res. & Dev. (HRDI-11)
Federal Highway Administration
6300 Georgetown Pike
McLean, VA 22101
Tel.: (202) 493-3147
Fax: (202) 493-3161
E-mail: terry.mitchell@fhwa.dot.gov

1 8 S U P E R PAV E M I X T U R E D E S I G N G U I D E

V I S I T

U S

O N

T H E

www.tfhrc.gov

FHWA-RD-01-052

W E B

A T :